Re: 4.10-rc1 on Nokia N900: regression, WARN_ON() omap_l3_smx.c

From: Pavel Machek
Date: Mon Jan 23 2017 - 18:33:28 EST


Hi!

> * Pavel Machek <pavel@xxxxxx> [161229 15:14]:
> > Hi!
> >
> > I forgot I had v4.10-rc1 running, and now I got warning on all the
> > consoles (hand-copied).
> >
> >
> > Unhandled fault: external abort on non-linefetch (0x1028) at
> > 0xfa0ab060
> > ...
> > Comm: kworker/0:0 Not tainted.
> > Workqueue: events musb_irq_work
> > ...
> > PC is at musb_default_readb().
> > ...
>
> This means the clocks are not enabled at that point.
>
> > WARNING: CPU: 0 ... at drivers/bus/omap_l3_smx.c:166
> > omap3_l3_app_irq+0xcc/...
> > Tainted: GDW.
>
> If you comment out postcore_initcall_sync(omap3_l3_init);
> in drivers/bus/omap_l3_smx.c you'll see the proper stack
> trace instead of the l3 interrupt trace. The system will
> hang at that point most likely.
>
> > I do have patches to allow nfsroot over usb. But they worked ok in
> > v4.9... Does anyone see it, too?
>
> Hmm not much has changed since v4.9. Are you sure you
> had v4.9 or some earlier v4.9-rc version?

It happened again in -rc5. I have full dmesg this time.

I started new thread with "

Subject: v4.9 to v4.10 regression: oops when USB cable is plugged in.

". It does no longer look completely scary.

Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Attachment: signature.asc
Description: Digital signature